Motifs in evolving cooperative networks look like protein structure networks

Abstract

The structure of networks can be characterized by the frequency of different subnetwork patterns found within them. Where these frequencies deviate from what would be expected in random networks they are termed "motifs" of the network. Interestingly it is often found that networks performing similar functions evidence similar motif frequencies. We present results from a motif analysis of networks produced by peer-to-peer protocols that support cooperation between evolving nodes. We were surprised to find that their motif profiles match closely protein structure networks. It is currently an open issue as to precisely why this is. © American Institute of Mathematical Sciences.